Business Process Modeling (BPM) explicitly presents the internal process sequences of a company, e. g. the creation of quotes and orders, and describes the “who”, “how” and “what” of the process activities. In the project ProPoneRe*, the working group Software Engineering (Prof. Gruhn) is investigating how modelers can avoid making mistakes or forgetting important aspects. Their approach: Together with researchers at the University of Koblenz-Landau and the West Saxon University of Applied Sciences Zwickau, they want to develop a recommending system that suggests suitable model elements during modeling and points out possible errors. 

The recommender system is to work on the basis of patterns from historical process models. Later, during modeling, these patterns are compared with the model fragments entered. To identify and categorize the patterns, the researchers analyze the labels in the historical process models using natural language processing (NLP). In addition, probabilistic methods (N-grams, probabilistic finite automatons) are to be adapted for working with process models in order to predict, for example, further process elements to be modeled or probable errors. 

In test person studies, the recommender system is evaluated on the basis of process models from industry and made available to the scientific community as a program library. 

*ProPoneRe (Support of Business Process Modeling by Model-Based Proposal Systems) is funded by the German Research Foundation DFG for a period of 24 months. Project partners of Prof. Dr. Volker Gruhn are Prof. Dr. Patrick Delfmann (University of Koblenz-Landau) and Prof. Dr. Ralf Laue (West Saxon University of Applied Sciences Zwickau)